Project Background Lung cancer represents a large burden of disease within the US and the VA population killing more Americans than the top three cancer types combined. Fortunately, the widespread adoption of CT screening for lung cancer is expected to result in more patients identified with early stage disease, however these patients still face a significant risk of lung cancer recurrence or development of a second primary lung cancer. National guidelines for follow up imaging surveillance exist, however structured data to evaluate guideline concordance and associations between guideline concordance and clinical outcomes are lacking. The current literature upon which the guidelines are based are not robust at the most recent data offer conflicting recommendations. Thus, the goal of this study is to examine patterns of post-treatment surveillance and determine concordance with national guidelines, impact of surveillance on clinical endpoints, and model best practices for optimal surveillance strategies in this high risk population. Project Objectives The objectives of this study are to (a) determine patterns of care and concordance with national guidelines for imaging surveillance, (b) determine the impact of guideline concordant routine imaging surveillance on clinical endpoints, and (c) evaluate the effectiveness of newer recommendations stratified by stage and cancer treatment on post-lung cancer treatment survival to clarify provider decision conflict. Project Methods To achieve these objectives, we will conduct a retrospective analysis of secondary clinical data linking patient records from multiple data sources. Patient demographics, provider information, inpatient ICD-9 codes for medical comorbidities, and diagnostic and treatment interventions from the Central Data Warehouse (CDW) will be linked to mortality records in CDW and other VHA vital status files. Cancer data will be obtained from the VA Central Cancer Registry (VACCR) including stage, treatment, and recurrence information. VACCR will also provide initial cohort identification. Raw CDW data files will be used for radiology text report information which are also available through CDW. To perform more efficient data collection, we will use a novel semi- automated chart abstraction method previously employed by our team for detection and categorization of lung nodules in an unscreened population. The method involves feeding electronic text imaging reports through a two-step process using coded data through SQL, a search tool to identify presence of potential key words related to each category, followed by manual chart review from highlighted text abstracts. Imaging tests will be categorized as to the indication for a given study, the presence or absence of recurrent or new disease as well as recommendations for follow-up. Patterns of care received will be compared to national standards to assess concordance with published guidelines and specific radiologist recommendations for follow-up. Clinical outcomes will be assessed to determine the association between receipt of guideline concordant care and detection of recurrent or new lung cancer, receipt of secondary therapy or interventions, and overall and lung cancer specific survival. We will then examine survival outcomes associated with adherence to recommendations based on the findings of the prospective randomized controlled Intergroupe Francophone de Cancerologie Thoracique (IFCT) group trial supporting less frequent surveillance imaging versus the newest National Comprehensive Cancer Center clinical guidelines supporting more frequent imaging for patients with late stage disease. Our goal in Aim 3 would be to test the impact of the newer recommendations that rely on tailoring imaging surveillance based on stage and treatment in order to reconcile the impending conflict between the two recommendations.